The TROPOspheric Monitoring Instrument (TROPOMI), due to its wide swath, performs observations over the ocean in every orbit, enhancing the monitoring capabilities of methane from space. In the short-wave–infrared (SWIR) spectral band ocean surfaces are dark except for the specific sun glint geometry, for which the specular reflectance detected by the satellite provides a signal that is high enough to retrieve methane with high accuracy and precision. In this study, we build upon the RemoTeC full-physics retrieval algorithm for land measurements, and we retrieve 4 years of methane concentrations over the ocean from TROPOMI. We fully assess the quality of the dataset by performing a validation using ground-based measurements of the Total Carbon Column Observing Network (TCCON) from near-ocean sites. The validation results in an agreement of $-\mathrm{0.5}\pm \mathrm{0.3}$ % ($-\mathrm{8.4}\pm \mathrm{6.3}$ ppb) for the mean bias and station-to-station variability, which show that glint measurements comply with the mission requirement of precision and accuracy below 1 %. Comparison to ocean measurements from the Greenhouse gases Observing SATellite (GOSAT) results in a bias of $-\mathrm{0.2}\pm \mathrm{0.9}$ % ($-\mathrm{4.4}\pm \mathrm{15.7}$ ppb), equivalent to the comparison of measurements over land. The full-physics algorithm simultaneously retrieves the amount of atmospheric methane and the physical scattering properties of the atmosphere from measurements in the near-infrared (NIR) and SWIR spectral bands. Based on the scattering properties of the atmosphere and ocean surface reflection we further validate retrievals over the ocean. Using the “upper-edge” method, we identify a set of ocean glint observations where scattering by aerosols and clouds can be ignored in the measurement simulation to investigate other possible error sources such as instrumental errors, radiometric inaccuracies or uncertainties related to spectroscopic absorption cross-sections. With this ensemble we evaluate the RemoTeC forward model via the validation of the total atmospheric oxygen (O₂) column retrieved from the O₂ A-band, as well as the consistency of XCH₄ retrievals using sub-bands from the SWIR band, which show a consistency within 1 %. We discard any instrumental and radiometric errors by a calibration of the O₂ absorption line strengths as suggested in the literature.

中文翻译：

---

应用于 TROPOMI 海洋太阳闪光测量的甲烷全物理反演评价

对流层监测仪（TROPOMI）由于其宽幅，在每个轨道上都对海洋进行观测，增强了对太空甲烷的监测能力。在短波红外 (SWIR) 光谱带中，海洋表面是黑暗的，除了特定的太阳闪烁几何形状，卫星检测到的镜面反射率提供的信号足够高，可以高精度地检索甲烷。在这项研究中，我们基于用于陆地测量的 RemoTeC 全物理反演算法，并从 TROPOMI 反演了海洋上 4 年的甲烷浓度。我们通过使用来自近海站点的总碳柱观测网络 (TCCON) 的地面测量来执行验证，从而全面评估数据集的质量。验证结果达成一致$-\mathrm{0.5}\pm \mathrm{0.3}$ % ($-\mathrm{8.4}\pm \mathrm{6.3}$ ppb) 的平均偏差和站到站变异性，这表明闪烁测量符合精度和准确度低于 1% 的任务要求。与温室气体观测卫星 (GOSAT) 海洋测量值的比较导致偏差$-\mathrm{0.2}\pm \mathrm{0.9}$ % ($-\mathrm{4.4}\pm \mathrm{15.7}$ ppb), 相当于陆地测量值的比较。全物理算法从近红外 (NIR) 和 SWIR 光谱带的测量中同时检索大气甲烷的量和大气的物理散射特性。基于大气和海洋表面反射的散射特性，我们进一步验证了海洋上的反演。使用“上边缘”方法，我们确定了一组海洋闪光观测，在测量模拟中可以忽略气溶胶和云的散射，以调查其他可能的误差源，例如仪器误差、辐射测量不准确或与光谱吸收交叉相关的不确定性-节。有了这个集合，我们通过验证大气总氧（O₂ ) 从 O ₂ A 波段反演的列，以及使用 SWIR 波段的子波段反演XCH _{4的一致性，显示一致性在 1 % 以内。我们通过校准文献中建议的 O 2}吸收线强度来丢弃任何仪器和辐射测量误差